1 Introduction

In this paper we investigate participation in online communities of people seeking behaviour change, such as weight loss or smoking cessation. On one hand, there is ample evidence that individuals benefit from various forms of health support exchanged in online settings (Maloney-Krichmar and Preece 2005; Preece 1999; Wicks et al. 2010). On the other hand, there are indications that individuals are typically reluctant to discuss their behaviour change online. People often fail in their attempt to change, and online contributions may be experienced as leading to potential embarrassment (Purpura et al. 2011). This is particularly prevalent on social network sites where people may be concerned about impressions given, sometimes inadvertently, to family, friends, co-workers and acquaintances (Morris et al. 2010; Newman et al. 2011). Even in anonymous online communities, the majority of contributions typically come from members who have successfully completed their behaviour change (Cobb et al. 2010; Maloney-Krichmar and Preece 2005).

We have characterised this kind of contradictory desire but also reluctance for online participation through the notion of the ambivalent socialiser (Ploderer et al. 2012). The stance may result from a deep ambivalence about the behaviour change itself: these individuals simultaneously want to keep up old habits but also desire a different, healthier life. Typically they prefer to read rather than actively contribute to online communities in the context of their behaviour change. Hence overall levels of interaction within groups of ambivalent socialisers are likely to be especially lean. Attempts to encourage personal contributions about behaviour change may be undesirable and counterproductive. Therefore, keeping users committed, attracting new users and encouraging contributions are ongoing challenges and many online communities go quiet (Kraut and Resnick 2012).

In this paper we report a design-based study to investigate the potential of an object-centred approach to facilitate online interaction among ambivalent socialisers. By object-centred approach, we mean the design of an online environment where social exchanges are, initially at least, mediated through objects of content rather than through direct person-to-person interaction. The study involved the design and evaluation of a mobile app to facilitate support among a group established to support smoking cessation. We set out to investigate the proposition that an object-centred approach to online participation would alleviate some of the concerns about making contributions in the context of behaviour change. The main premise is that object-centred interaction allows participants to engage while retaining initial anonymity and control over their exposure to the group. Interaction is shifted from direct personal exchange to object-mediated exchange. An object-centred approach can allow social interaction, in various forms, but without demanding it as the basic mechanism of participating.

The focus of our study was not on cessation outcomes, but rather on the potential for an object-centred design to engage ambivalent socialisers, and also to characterise the emerging forms of online object-centred sociality. The concept of ‘object-centred sociality’ in online settings has been explored by Engeström (2005) and is associated with metatheoretical approaches, principally activity theory (Kaptelinin and Nardi 2006) but also others like actor-network theory (Latour 2005) and sociomateriality (Suchman 2005, 2007). Engeström’s work was particularly influenced by the work of Knorr-Cetina and her observations of object-centred sociality in contemporary Western societies (Knorr Cetina 1997). In light of the decreased importance of relations with traditional communities and the increased importance of knowledge in such societies, Knorr Cetina states that ‘object relations substitute for and become constitutive of social relations’ (Knorr Cetina 1997, p. 9). Engeström argues that objects play a similar role in online settings, implying that successful social network sites not only bring people together but also provide them with objects that mediate relationships between individuals. Photos on Flickr, videos on YouTube, and a variety of media on Tumblr all highlight how shared objects of interest mediate online social interaction on a large scale.

The notion of object-centred sociality has been applied to a variety of works in HCI and CSCW. Hagen and Robertson (2010) have explored the opportunities of object-centred sociality via social technologies for participatory design work. Karnik and colleagues (2013) applied object-centred sociality to media sharing on social network sites, studying how people establish relationships based on these object-centred interactions and the gratifications they derive from it. Finally, Comber and Thieme’s (2012) BinCam system used an object-centred approach to behaviour change, to raise awareness about food waste and to improve recycling in shared households. The BinCam system captured and shared images of disposed waste on an online social network to stimulate interaction about recycling and food waste.

Hence, the aim of the study reported here was to build on these concepts and developments of object-centred sociality and to apply them to our chosen domain of smoking cessation. Our intention was to explore ways to encourage productive online interactions between people attempting to quit smoking. The study was driven by two questions. Firstly, we asked a general design question of Whether and how an object-based design supports interaction between ambivalent socialisers. Particular emphasis was put on how the object-based interface might allow participants to interact while controlling their engagement and exposure to the group. Secondly, we explored a related theoretical question of What is the nature of the social interactions that arise around an object-based online community interface? We took Preece and Schneiderman’s (2009) reader-to-leader framework as a starting point because it offers a practical and theoretical view of participation in an online community. In this view there are typically a large number of readers, also sometimes known as lurkers, who consume content but make no active contributions themselves. The aim of community designers and hosts is to encourage these readers to become active contributors of content, and then later possibly collaborators who work with other members, or even leaders who influence directions for the group as a whole. Hence, the second question was exploring if the reader-to-leader framework would provide an adequate account of participants’ involvement in a behaviour change support group. Considering that interaction amongst ambivalent socialisers is likely to be lean, how else might the nature of interaction be interpreted and understood?

Section 2 provides an overview of the context of smoking cessation and the object-centred application that we developed. Section 3 describes the field study design. In Section 4, we report the observed forms of object-centred social interaction and offer a characterisation of them to complement the reader-to-leader framework. Section 5 discusses how our findings enhance current understanding of online interaction with and through objects. We conclude with a discussion of the applicability of an object-centred approach to other areas of online participation within the domain of behaviour change and elsewhere.

2 Study context and system design

2.1 Study context: smoking cessation

Smoking cessation is a suitable domain to study the potential of object-centred approaches to facilitate interaction between ambivalent socialisers. Research with smokers in Canada, the United States, the United Kingdom, and Australia suggests that around 40 % of smokers attempt to quit in a given year, yet the vast majority (97 %) of these attempts fail within one year (Borland et al. 2012). Consistent with the above definition of ambivalent socialisers, it is known that social media like Facebook hold great potential benefits for quitters (Cobb et al. 2011), but people are typically reluctant to raise their behaviour change online before they feel they have succeeded (Newman et al. 2011). Studies show that reluctance to contribute to online groups often stems from the perceived risk of failing to quit and the potential embarrassment associated with failure (Burri et al. 2006; Cobb et al. 2010).

A key challenge in staying quit is managing cravings for cigarettes, particularly in the first month after quitting (Herd et al. 2009). During this time quitters frequently experience cravings, often triggered by particular times of the day, activities, places, emotional states, and other smokers. Nicotine replacement products can reduce withdrawal symptoms, but people have to learn how to manage these situations and cope with cravings to prevent a relapse (Shiffman et al. 2005). Popular strategies to cope with such cravings are keeping busy, breathing exercises, food and drink as well as self-encouragement (Shiffman et al. 2005). Although ex-smokers report that they rarely get cravings after 2 to 4 weeks, they still need to remain mindful of their coping strategies when temptations to smoke present themselves.

2.2 The DistractMe app: design and content management

To investigate the object-centred approach to problem online participation for ambivalent socialisers, we developed and evaluated a smartphone app for quitters called “DistractMe”. To follow an object-centred approach, the first step was to identify a kind of digital content object that participants could usefully employ in their own personal quit attempt without any necessity to interact with others. Around this object-based and solitary task, the potential for incident social interaction would be built.

For this, we designed a tool intended to help smokers distract themselves during cravings to smoke. Users of the app could compile a suite of digital content objects (videos, games, websites, images) that they thoughts would help to distract themselves when needed (see Figure 1). Distractions in general are an effective means to divert people’s attention from cravings (Shiffman et al. 2005).

Figure 1
figure 1

Key features of the DistractMe app: a a scrollable list of distractions and b tips to cope with cravings with the option to add a comment. The red circle in the navigation bar highlights notifications of responses from other users.

Full size image

DistractMe was intended for personal use to cope with cravings, but it also created possibilities for social awareness and interaction around the exchange of distraction content, through an online environment in which objects could be shared, commented on, and rated as favourites. We also included social traces like the number of views, comments and favourites to each distraction (see Figure 1).

Further, we added a notifications feature to highlight responses from other users and thereby encourage contributions and comments. As discussed by Engeström (2005) and as evidenced by popular social media like YouTube and Pinterest, distractions are also a popular means to encourage online participation.

We incorporated in the app’s design two categories of content objects to support quitters with potentially different affordances for social interaction. In the first category were pure distractions, as described above. In the second category, labelled tips, were explicit statements about how to approach a quit attempt, such as ‘drink water during a craving’ (see Figure 1). Tips were still strictly objects of content, but their nature as textual pieces of advice meant that they were possibly more overtly socialisable objects. We envisaged different uses of distractions and tips. Distractions were to serve situations in which participants wanted to avoid thinking about smoking, while tips were to serve situations where participants wanted to think about and fortify their quit attempt. Distractions also allowed for interaction largely independent of any reference to the issue, especially in relation to the person’s progress or to challenges they are facing.

Given that users are typically reluctant to participate in support groups for behaviour change, we sought to protect their privacy in a number of ways. Users did not have to sign in to the app to access content, unless they wanted to add content or comments. They could freely choose a user name when they signed up, but the system did not offer profile pages with photos, personal information or information about a user’s quit attempt. Users could choose to share content without sharing their user name, and they could add content for their personal use only. Each contribution was immediately visible to the contributor, but required approval from a moderator to be visible to other users.

The app was not available in the app store but only through a download from a specific website. This helped to protect the privacy of the support group, because only users who consented to take part in our study had access to its content. However, this was also a limitation, because it resulted in a smaller population of users and it was unlikely that the community input would be enough to be self-sustaining. Hence the researchers played an active role in adding content to the app. With the help of our research partner Quit Victoria we prepared 171 distractions and 179 tips. Approximately half of these tips and distractions were available from the start of the study and half of them added throughout the field study. Tips were added under the user name Quit Victoria. Distractions were added under name of the first author, though most of them anonymously to avoid dominating the updates. The first author also added 37 comments on tips and distractions to seed group activity.

3 Method

We conducted a field study with the DistractMe app with 18 smokers who were trying to quit smoking. Observations of their activities, interactions and experiences were made through interviews, diaries and log data of online activity.

3.1 Study participants

We recruited 28 participants to try out the DistractMe app during their quit attempt. Participants were recruited through Quit Victoria’s telephone counselling service and their Facebook and Twitter channels as well as through the University of Melbourne’s staff and student mailing lists. We selected only participants who were planning to quit smoking in the month following the first interview or had quit no more than 1 week before the interview. Furthermore, they needed to own an iPhone to run the DistractMe app. All participants were ambivalent socialisers who used social network sites like Facebook and Twitter but were reluctant to discuss their current quit attempt on these sites.

The findings presented here are based on 18 active participants. The other 10 were excluded because log data showed that they did not use the app at all apart from initial browsing. Further, they stated that they had not made a quit attempt with the app, or that they found a different method to quit.

The 18 active participants accessed the app at least three times after the first interview. The cohort was diverse in terms of age (ranging from 19 to 53 years, average age of 33 years), gender (13 female, 5 male) and their commitment and success in stopping to smoke cigarettes (15 made a quit attempt, 7 of them described themselves as non-smokers at the time of the second interview). As summarised in Table 1, they varied in their contributions to the DistractMe app. As expected with ambivalent socialisers, the level of participation was low with an average of 2 distractions and 1 tip posted per person during the 6–12 weeks that they had the app. Despite their ambivalence, however, all 18 participants made at least one publicly visible contribution to the app, either through a comment, tip, distraction, or by adding existing content to their favourites.

Table 1 Contributions to the DistractMe app during the field study by the 18 active participants.
Full size table

3.2 Data collection

As summarised in Table 2, the findings presented in this article are based on two interviews per participant, diaries and log data. At the start of the study (November and December 2012), the first author interviewed each participant about their plans to quit and their usage of iPhone apps. During this interview, participants downloaded and familiarized themselves with DistractMe. The interviews were conducted face-to-face (21) or via telephone (7) and lasted approximately 60 min. The participants received $25 per interview to contribute towards travel expenses and phone data charges.

Table 2 Methods used during the field study.
Full size table

The second round of interviews was conducted in January and February 2013, because most participants planned to quit around New Year’s Day and they expected to use the app during the first few weeks of January to cope with cravings. The interviews focussed on their engagement with the app to cope with cravings, the various sources of support and distraction they received, and their interactions with other users. These interviews lasted between 30 and 60 min, depending on the participant’s level of engagement with the app.

Log data was collected throughout, recording physical interactions with the app concerning the posting of tips, distractions and comments, as well as objects accessed. We also asked the participants to diarise their experiences with the app once a week during the trial to get qualitative feedback at the time of their quit attempt. The diary entries provided contextual information that helped us interpret the log data throughout the study. Furthermore, it provided insight into the participants’ progress in their quit attempts.

3.3 Data analysis

From the first interview onwards we wrote analytic memos for each participant that contained their data (posts to the app, log data, as well as their diary entries), observations during the interview and interpretations of their online activities (Birks et al. 2008). Whilst the analysis was sensitised by the concepts of the reader-to-leader framework of online participation (Preece and Shneiderman 2009), we also anticipated gaps in that framework in this context, and looked for new ways to characterise the forms of object-based sociality that arose. For example, very early during the field trial it became evident that some of the participants posted tips that disclosed extensive personal information that sat somewhere between the categories of contributing and collaborating. Furthermore, diary entries hinted at different experiences around simple acts of reading content. We used the second interview to discuss and check such observations with the participants. At the end of the field trial we had a variety memos that captured our observations and preliminary assertions about different forms of engagement. The first author used NVivo to code all qualitative data (interview transcripts, diary entries, posts to the app) following Miles and Huberman (1994), and these were then refined and structured into themes, as illustrated in Figure 2.

Figure 2
figure 2

Three ways to describe participants’ interaction through the DistractMe app. Inside the circle shows the physical interface-level actions taken on objects of content. The outer layer shows two abstractions: participation in the community in terms of Preece and Schneiderman’s reader-to-leader framework (above the circle); and the present article’s alternative focus on acts of varying sociality afforded by the app’s object-mediated interaction (below the circle).

Full size image

4 Findings

The key thematic concepts that emerged through the analysis are depicted in Figure 2, and are now summarised in advance of the detailed findings. Inside the circle of Figure 2 are the terms used to describe the interface or physical level interaction inherent in the DistractMe app design. There were three kinds of content objects: distractions, tips, and comments attached to them. These were subject to four basic actions. Consuming implies reading, viewing or listening to a particular item of content, as distinct from browsing through the list of content objects as a whole. Other actions taken by participants were posting new objects of content, and classifying tips and distractions by, for example, tagging them as ‘favourites’.

Outside the circle in Figure 2 are two alternative ways to conceptualise the meaning of participants’ online interaction. Above the circle, are the stages of the reader-to-leader framework that were adapted here to accommodate the lean interaction of the smoking cessation group. We used the framework to refer to activities (like reading) rather than kinds of users (such as readers), because most of the participants moved back and forth between the various activities. Reading, in this context, was taken to imply solitary browsing and consuming of content. Contributing implied posting new content in some form, while collaboration was taken to be acts of intentionally soliciting or offering support for each other’s quit attempts. Leading would imply activities such as introducing new users and moderating content. However, we did not observe any such activities, partly because leadership was simply not required and partly because even the most active participants stopped engaging with the app once quitting had ceased to be an issue.

Below the circle in Figure 2 we propose an alternative way to conceptualise participant interactions that emerged through our analysis. Here we identify a dimension of sociality to capture the way participants experienced the presence of the social group, as opposed to the reader-to-leader framework’s emphasis on objective levels of participation in a community. At one end of the spectrum of sociality, there were instrumental actions that dealt directly with objects without any experience of the people who posted them. This is associated with solitary use of the DistractMe app as a personal tool to assist with a quit attempt. Moving away from this non-social instrumental use, users might in contrast experience various forms of sociality by engaging with the social presences of others in the group. The first form of sociality was simply to be receptive to the presence of others. Another form we described as covert, referring to where participants experience their actions as influencing the group but without expressing an individual presence. In contrast, this became expressive sociality, typically through an act of disclosure, if participants explicitly referenced themselves and their situation to the group. Going further still, the experience became mutual sociality if participants explicitly referenced the existence of others in the communal space.

This dimension of sociality is not intended to negate the reader-to-leader framework. Rather, the emphasis on sociality offers a complementary view of online interaction around the object-based interface and its use to support ambivalent and lean interaction as commonly found in many behaviour change domains. In the following sections we organise our findings around the new forms of engagement identified from the non-social instrumental actions through the receptive, covert, expressive and mutual form of sociality.

4.1 Instrumental actions

We begin by considering how participants used DistractMe in a purely instrumental manner as a personal tool to assist with a quit attempt. As we will report, this type of engagement involved consuming content as objects to cope with cravings, contributions of objects without consideration or awareness of the people who posted them.

4.1.1 Instrumental consumption

Only three participants reported episodes where they used the distraction objects in the intended way to take their mind off cravings. For example, participant 2 was confronted with a stressful situation at work, which almost made her fall back into smoking to cope. However, she used the DistractMe app in concert with other apps to find distraction and to wind down without relying on cigarettes. I came home thinking, you know, ‘I don’t know what to do with myself, I don’t know how to process all this.’ And after I’d looked at a few newspapers online and had a bit of an interaction with some social media, I was sliding across the screen and went, ‘Ooh, there we go, bingo!’ I logged on and all of a sudden 20 min later sort of went, ‘Well,’ and because I had been thinking I want to come home and have a cigarettes really badly, but I didn’t. And so after about 20 min the craving was gone and I didn’t even think about cigarettes for the rest of the night and felt a great deal more settled.” (Participant 2, interview 2)

In an unintended use of the app, two participants used the tips to cope at the moment of suffering cravings for cigarettes. In place of distractions, they used the tips to continue to fortify their resolve to stay quit. For example, participant 17 used the videos and graphics in the tips section to remind herself of the health risks associated with smoking. “If I’m stressed and if I feel like I have to run to the 7-Eleven [and buy cigarettes], I just feel like OK, let’s watch some creepy information about smoking on the application.” (Participant 17, interview 2)

4.1.2 Object-centred contributions and responses

The majority of distractions and tips were instrumental or purely object-centred contributions, in that they contained or linked to relevant content but lacked personal references including any to the potential audience. Twenty seven of the 37 distractions posted fell into this category. They typically offered to provide distraction through entertainment, through humour (15) like funny animal photos and videos (see Figure 3a), infotainment (7) like videos of nature and foreign countries, websites that offer similar distractions (4), or games (1).

Figure 3
figure 3

a Example of distraction linking to series of animal photos. b Example of a personal story that yielded strong responses from the participants.

Full size image

Similarly, approximately half of all comments posted on the app were only indirectly social in that they responded primarily to the shared object (distraction or tip) without a personal or interpersonal reference. For example, 15 out of the 21 comments on distractions were merely expressions that the participants liked the content, such as “Love it!” (Participant 17), “Ha-ha, that’s funny” (Participant 15) or “Such a good show!” (Participant 11). The remaining comments expanded on the distraction either by providing additional information (e.g., about the source or the actor of a video) or by engaging with it in a playful manner. But in all these cases, the act of commenting seemed almost instrumental with the focus on the object itself. We observed similar object-centred responses on tips, yet to a lesser extent. Only 8 of the 28 comments on tips were brief contributions to say that they liked or recommended the content, such as “great tip” (Participant 19) and “so true” (Participant 17).

Such comments are perhaps best regarded as contributions rather than collaborations because no attempt was made to connect with the person behind the objects. Yet they border on the notion of collaboration in that they potentially involve one person responding specifically to another person in a way that could advance their goal of quitting. This ambiguity highlights the power of the object-based design approach. The commenter acted in a way that protected exposure to the group, yet may offer a form of social support. And in some cases, these responses were a stepping-stone towards other forms of sociality.

4.2 Receptive sociality

The simplest and least demanding form of sociality was simply to be receptive to the presence of others users. This was manifest in such actions as looking for traces of social interaction as well as consuming content in which are other users were explicitly inscribed; this was mainly stories of other users’ experiences with quitting that were shared through the tips section of the app.

4.2.1 Receptive sociality in browsing distractions, tips and comments

As intended in the design of the app, at those times when participants were not suffering from craving or adverse emotions, they could plan their quit attempt by browsing and reviewing distractions on the app for later use: “I used it sitting at home, when I haven’t necessarily been wanting a cigarette, just to see what was on there.” (Participant 8, interview 2) Also as intended, most participants engaged with the tips in their downtime to fortify their resolve for the quit attempt. Typically, participants used the information in the tips for self-encouragement or to remind themselves of the negative consequences of smoking.

While such browsing could be purely instrumental, at other times it could involve heightened attention to the other quitters who had posted the content; and so involved what we are calling receptive sociality (Figure 2). One indication of this was the way browsing was often joined with existing socially-oriented routines of checking emails and social media: “When I’m bored, I just go through what do I have to do on my phone, go check Facebook five times. I go through my photos, and then I’m going checking something on DistractMe, which is fun.” (Participant 15, interview 2)

When participants browsed through comments attached to the distractions and tips, their activity was highly receptive to the presence of others. Some participants reported that at times they entered the app purely to look for new comments from other users, without necessarily contributing objects, interacting with other people, or further engaging with the content: “I occupied myself with the app a few times, but now that I’ve seen a lot of the distractions I’m most often just looking to see if anyone has made a comment in there.” (Participant 1, diary)

4.2.2 Receptive consumption

For some kinds of content, this receptive sociality grew to levels that some participants described in terms of empathy with the people who posted the content. Participants attributed enhanced value to objects that they experienced as social postings. For example, participant 2 (interview 2) stated: “I liked personal hints. I think if someone’s taken the time to put something personal up there, then I always felt that would be of more value.”

The participants typically empathised with tips that conveyed difficulties as well as small successes in the quit process. While the tips provided by Quit Victoria provided factual information and advice for coping with such difficult situations, the participants engaged with tips that simply acknowledged that quitting is hard, because it showed them that they were not by themselves with their struggles. In light of the difficulties of quitting, some tips celebrated small successes like making a quit attempt or staying quit for one day, something that many participants could relate to but found difficult to convey to non-smoking partners and friends. For example, participant 8 appreciated the tip displayed in Figure 3b, which started with “the first day was very hard” before discussing tips that got that person through the first few days.

“I really liked the personal stories. I said last time that I didn’t really care about what other people said on the app, but there are stories, like people saying how long it’s been . . . When someone hasn’t had a cigarette for 3 days and they were talking about how difficult that was and how good they felt after the first day, which is really true. If you are smoking every day and you can go a day without one, it’s quite a big achievement. And then the second day again, you go through a similar sort of thing. So that helps. It was good to see that those early stages are very similar for those people.” (Participant 8, interview 2)

4.3 Covert sociality

While the majority of distractions and tips lacked personal references, including any to the potential audience, posting distractions was not always a purely instrumental act. In many cases it was also clearly associated with the intention of affecting others in the group but without declaring the existence of oneself. We identified these acts as covert sociality. For example, participant 17 posted impersonal motivational messages in the distraction section of the app, like “every accomplishment starts with the decision to try.” The object-based design afforded an easy transition into this covert experience, allowing them to act socially while not requiring users to put forward a representation of themselves, not even an anonymous or symbolic one. Posting distractions provided an easy stepping-stone from reading towards contributing, in that it allowed the participants to contribute objects that were impersonal and relevant to quitting, without having to reveal any information about themselves or their experience. Moreover, distractions were easy to come by via social media and recommendations from personal networks. Distractions could be posted without having a close understanding of the audience.

One relatively extreme case of covert sociality concerned participant 20 who rarely utilised DistractMe content for herself because she quickly relapsed in her quit attempt. Nevertheless, she continued to post distractions to help others, assuming that they would appeal to most other people. “If I’m studying at night, I usually try and stop myself about once an hour just to give my brain a break for about ten minutes, and what I’ll do is I’ll go into YouTube. Most of my friends and family know the sort of stuff I like, and they’re constantly sending me links . . . A lot of my posts [on DistractMe] were animal clips, but they were the ones that I think could appeal to just about anybody because they were so good.” (Participant 20, posted on DistractMe)

Like distractions, tips also allowed participants to be covert and share information without having to talk about their personal quitting experiences. In fact, only 7 of the 21 tips were written in a pure object-centred way that lacked any sign of sociality. These tips gave suggestions to cope with cravings like keeping hydrated, or sometimes links to books and other apps for smoking cessation. However, as we will discuss next, most tips went further and also disclosed personal information about the creator.

4.4 Expressive sociality

While covert acts involved a user wanting to affect the group without assuming any presence, a different experience of sociality was manifest when self-references were made, even if in an anonymous or symbolic form. We describe this as expressive sociality. This typically took the form of self-disclosure and was visible in contributions of content that referred to the existence of the poster. Such self-disclosure was more common with tips than with distractions, because tips were typically textual and composed by the participants, whereas all distractions linked to existing online content. Only one distraction linking to a popular news website was framed through a caption in a personal way, stating “Reading the news helps!” (Participant 2, posted on DistractMe). However, approximately half of the tips (10 out of 21) contained some form of previously unknown personal information about the participants. Unlike the contributions discussed above, these tips here disclosed personal experiences from previous quit attempts as well as their ongoing progress. Some posts contained simple personal suggestions to cope with cravings or to avoid them altogether, such as “I’ve given up alcohol for a few weeks as I think it might help me with giving up. Alcohol & cigarettes have formed such a close bond in my head that I think this might be a way to break that bond.” (Participant 2, posted on DistractMe)

Three participants went even further and talked about the status of their current quit attempt, including reflections on techniques used and the resulting experience. For example:

“Day 6, still going strong. Thrown out all my rolling paraphernalia, my girl friend is being really supportive. My friends on the other hand aren’t taking me seriously. Been smoking for 5 years, it’s hard to imagine myself as a non smoker, I feel like its a part of my identity, which is pretty scary. I’m doing my best to distance myself from my friends that smoke so I’m not tempted, but I can’t do that forever. I have found having something in my hand helps DistractMe, my current toys are toothpicks and chewy. Another really helpful quit app that compliments (I think) this app is “since I quit” free on the App Store and tells you; how longs its been since you smoked, how much money you have saved, how many cigarettes you haven’t smoked, and how much you have extended your life by not smoking. Check it out! Keep going strong!” (Participant 24, posted on DistractMe)

The extent of self-disclosure within such posts, particularly statements of progress in the quit attempt, was noteworthy given reported concerns about privacy, peer pressure and potential embarrassment due to failure to quit. In line with previous research we also found that many participants were indeed reluctant to discuss their quit attempt with other people due to fear of potential failure and embarrassment. For example, participant 11 (interview 2) stated: “I appreciate them sharing, but I’m kind of guarded about things like that, and especially quitting smoking is hard. I didn’t want to say anything to a lot of my friends because I knew I was going to start smoking again. I didn’t want them to know that I failed at it.”

4.5 Mutual sociality

So far we have described how interactions with content in the app could vary in its associated experience of sociality from instrumental object-based acts, to receptive, covert or self-expressive acts. A fifth kind of experience was also observed that we describe as mutual sociality, referring to acts where a participant referenced the presence of others in the group.

4.5.1 Invitations to collaborate

Some participants articulated their tips and distractions in ways that explicitly reached out to others and invited them to collaborate on their quit attempts. These included but went beyond self-disclosure, and explicitly addressed the imagined audience through encouragement, suggestions to try out particular distractions, and questions. These posts were one-way offerings to start a conversation and to collaborate, though not all of them received an explicit response.

Six distractions contained such references to an audience. For example, participant 24 posted a link to a website about gadgets with the caption: “Thank me later, great distractions on here.” In a subtle way this caption addressed other users of the app and reinforced a group goal of sharing distractions in order to stay quit. Participant 17 posted four distractions that crossed over with tips, as they all encouraged others to stop procrastinating. For example, one of the slogans stated “Every accomplishment starts with the decision to try. Try today!” These posts show that distractions offered opportunities for inviting collaborative action, even though the majority of distractions lacked them.

As with self-disclosure, invitations were more visible through tips because they were often self-composed text messages. Six tips contained explicit references to the audience, such as the following suggestion to use money as a motivation to stay quit. “Reward yourself! Been less than a week and I already have saved roughly $25 by not smoking. That money’s going towards Xmas shopping and junk food! Remember to reward yourself, it’s great motivation to think of all the money your saving.” (Participant 24, posted on DistractMe) Weaved into this object-based interaction, are references to the author and the reader that create an experience of mutual sociality.

4.5.2 Mutual sociality through reciprocal self-disclosure

More than half of the comments on tips (17 out of 28) contained some form of personal information about the person responding, typically in response to other people’s personal stories. As discussed in Section 4.4, many participants engaged more positively with tips when they could get a sense of the person behind them. The participants also expressed that reading other people’s stories in turn also encouraged them to share their own personal comments: “I love reading everyone’s day to day dealings with quitting, it makes me want to contribute as much as possible.” (Participant 24, diary) The following example shows how a personal tip by participant 2 led to a response by participant 11 where she empathised with participant 2, and beyond that, disclosed information about her current quit attempt.

Participant 2: “Step Away From That Cigarette! I like to “tidy” (read obsessively clean!) when I feel the urge to smoke. It occupies my mind & distracts me completely from wanting to light up. And the bonus is that you end up with a lovely, sweet smelling house!”

Participant 11: “Today is my 3rd day with no smokes, last night I rearranged my bookshelves. You are not alone!”

This example shows how through the reciprocity of tips and comments, self-disclosure sometimes facilitated social interaction between individuals about a shared goal or, in other words, collaboration. Distractions, on the other hand, could not support this pattern, lacking personal information in the initial posts and therefore also in the comments on them.

4.5.3 Mutually supportive posts

Some comments on tips went beyond personal information and contained an explicit attempt to help another person to stay quit. These comments offered support through positive encouragement and thereby represent collaboration to quit smoking. For example, participant 8 responded to a story that was posted anonymously showing empathy, disclosing personal information about his status, and offering encouragement to stay on track. The word ‘we’ highlights the mutual nature of this comment.

Anonymous tip: “I’m half way into day 4 and it is great. The first day was very hard, with the help of the patches in the first 24 hours was good, since then I have been using the microtabs, the 2nd day I used 4 tabs, the third day I used 3 and today I have only used one; also cut back on coffee and have started drinking water instead. Had my first call back today from the Quitline and it was very helpful. All the best to you all!”

Participant 8: “Congrats! Just started day 3 myself, on patches. Used gum as well the first 2 mornings but didn’t wake up dying for one today! Keep at it, we’ll get there!”

Overall, we observed only five instances of explicitly articulated mutual support around tips and none around distractions. The scarcity of explicitly supportive reflects the limited number of participants in this study as well as the stance of the ambivalent socialisers that inhabit the smoking cessation domain. It must also be noted that several participants regarded their contributions of objects as acts of support even though they lacked personal and interpersonal references.

5 Discussion

We began this paper with the problem of facilitating social interaction between people attempting to give up smoking. We identified this population as ambivalent socialisers in that they might want and benefit from social support to quit, yet they are reluctant to interact with others due to concerns about failure and potential embarrassment (Newman et al. 2011; Purpura et al. 2011). As with other behaviour change domains, designing online spaces to provide the kinds of social support that people want and/or find useful is difficult.

The DistractMe app reported here was designed and evaluated to test an object-based approach to this problem. The design intention was to construct an online space in which quitters primarily interact with objects of content for their own personal use. For this, they were provided with objects that could serve as distractions during cravings for a cigarette and objects that contained tips on how to give up smoking. Around the personal use of these digital objects, participants could, secondarily, interact with each other by sharing and commenting on content objects. What was secondary for the users, was the primary concern for the researchers who set out to explore the different forms of sociality that arise through this object-based design.

The findings have shown that users prefer to interact with the app by browsing and consuming content to help them stay quit. Publicly visible interactions with other users through contributed content and collaboration happened to a lesser extent, as predicted by the reader-to-leader framework (Preece and Shneiderman 2009). Within all of these interactions we observed different forms of sociality, which captured the participants’ experience of the presence of the social group through objects available on the app. These forms of sociality ranged from purely instrumental sociality with the objects available on the app, to subtle engagement with other users through receptive and covert sociality, to explicit interaction with other users through expressive and mutual interactions. As illustrated in Figure 2, these forms of sociality provided an alternative to the reader-to-leader framework (Preece and Shneiderman 2009) to account for the nuanced interactions through an object-centred system. The significance of this alternative characterisation of sociality as well as the object-centred design approach will be discussed in more detail in the next two sections to answer our research questions.

5.1 Designing for interaction around objects

The first question was whether the object-based design would support interaction between the ambivalent socialisers of the smoking cessation group. The findings showed that the object-centred approach successfully encouraged different forms of interaction. As predicted by the reader-to-leader framework (Preece and Shneiderman 2009), reading was a popular form of participation because it allowed engagement with behaviour change without exposing oneself to potential peer pressure and embarrassment due to failure to change (Purpura et al. 2011). For these reasons, previous research on Facebook (Newman et al. 2011) and anonymous online communities (Cobb et al. 2010) also suggest that people are reluctant to contribute before they have successfully changed. In contrast to these studies, however, most of the participants in this study made publicly visible contributions at very early stages of their behaviour change while they were still likely to relapse. In fact some participants continued to contribute even after a relapse. This was facilitated by the DistractMe app’s focus on objects as a stage on which online social interaction could occur, rather focussing on the participants’ personal experiences with behaviour change. While such object-centred contributions have worked well in other settings of online participation, such as music and photography (Engeström 2005; Karnik et al. 2013), this study showed that the approach also facilitates contributions in a personally sensitive setting of health behaviour change.

Beyond merely demonstrating that objects facilitate participation, the findings highlighted how the two categories of objects in the DistractMe app—distractions and tips—influenced participation at the different stages. Both kinds of objects supported reading, contributing and collaboration, however in different ways.

The first difference concerned how these objects were created, with implications for the contributions to the app as well as for collaboration between users. Distractions were typically references to external content like YouTube, Reddit, Tumblr, etc. Tips also sometimes linked to external content, but the majority of tips were composed for the specific audience of the DistractMe app. While the referencing and repurposing content from other social media and news media lowered the effort of contributing and led to a larger number of contributed distractions, they also generally decreased opportunities for self-disclosure and invitations for further interaction. These kinds of contributions were more visible through tips. As pointed out by the literature on online communities, rhetorical strategies found in the tips such as disclosing personal information and asking questions elicit responses from other community members and thereby foster collaboration (Burke et al. 2007). In summary, there is a trade-off in the content creation for object-centred online communities. Being able to easily reference existing content facilitates contributions, whereas composing content requires more effort but fosters collaboration.

A second difference lay in how these objects directed the attention of the users away from or towards the app, which was important for both reading and collaborating. The role of distractions was to take people’s minds off cravings, as suggested by findings on relapse prevention strategy (Shiffman et al. 2005). In doing so, they directed users to content on YouTube, websites, games, etc., and ultimately away from the app and from interacting with other users of the app. In other words, the purpose of distractions as a coping mechanism stood in conflict with their purpose as mediating objects between users. Tips on the other hand maintained a focus on the behaviour change, which drew people into giving more attention to the app and so encouraged personal interaction with objects as well as interpersonal interaction through objects.

Finally, distractions and tips differed in the ways in which they encouraged reading. Distractions on the app were often browsed, but received mixed feedback from the participants in terms of their utility, because the number of updates was rather low when compared with millions of posts on popular social media like YouTube, Reddit or Pinterest. As a result, distractions on the DistractMe app lacked the novelty and potential surprise that other social media offer as well as the opportunity to tailor distractions to personal interests. Furthermore, DistractMe lacked the personal context that sites like Facebook offer (Joinson 2008), where even a low number of distractions or posts unrelated to one’s personal interests are given meaning by the fact that they were posted by a friend, family or close contact. Tips on the other hand, received their relevance through the shared experience of quitting amongst the users of the app. Therefore, any kind of personal information included in the tips enhanced their value because they offered an opportunity to empathise with other users (Preece 1999). As shown in this study, tips with personal information were used to fortify the quit attempt by firming up one’s resolve to cease smoking. Beyond that, these tips invited interactions centred on both objects and the individuals behind them, because the participants reciprocated the benefits that they received from objects enriched with personal information.

5.2 The nature of interactions arising around object-based systems

The second question investigated in the study concerned the nature of social interactions that arise around an object-based interface. Engeström’s (2005) discussion of object-centred sociality as well as more current examples of social media structured around shared objects like YouTube, Reddit or Pinterest suggest that this approach can lead to highly engaging experiences. The findings of this study, on the other hand, show that at least in its early stages the online community provided a rather lean form of sociality. The experiences of its members were primarily concerned with instrumental and receptive consumption of the objects, whereas self-disclosure and mutual support were still limited.

As expected, social interaction around the objects in DistractMe was lean in because of the small size of the study cohort and the participants’ ambivalence about discussing their behaviour change (Ploderer et al. 2012). A longer study with a larger community would partially address this limitation, because it would increase the amount of concurrent use amongst the participants and may lead to a larger number of interactions. However, it would not address the participants’ ambivalence about giving up smoking, their concerns about failing in their quit attempt, and their subsequent reluctance to disclose personal progress. Participants in this study primarily engaged with the app after quitting, but abandoned DistractMe when they relapsed. Furthermore, the findings showed that also participants who successfully quit smoking stopped interacting with other users of the app. Unlike in other much larger and sustained online communities in this domain, where successful quitters support those who are currently trying to quit (Burri et al. 2006; Cobb et al. 2010), the successful quitters in this study were at a stage where they felt not yet safe enough and sought to avoid any reminders of smoking. While in some online communities long-term members become latent users who decrease their content contributions but nevertheless stay committed to the community (Velasquez et al. 2014), the participants in this study abandoned the support group when their withdrawal symptoms subsided and DistractMe had become obsolete.

The leanness of the social interaction meant that Preece and Shneiderman’s (2009) reader-to-leader framework provided only a broad-brushed account of the participants’ involvement in the online smoking cessation group. In terms of the reader-to-leader framework, their interactions were predominantly concerned with reading, with limited contributions and collaborations. In our findings (summarised in Figure 2) we have provided an alternative, more nuanced view of the participants’ sociality, which characterizes the nature of their experience of the social group as mediated through the objects available on DistractMe.

At one end of the spectrum of sociality, we have discussed instrumental actions that engage primarily with objects rather than the individuals who posted them. This form of sociality was visible in the solitary use of DistractMe as a tool for browsing and consuming of tips and distractions in order to assist with quitting. In this mode, participants contributed content for self-distraction and added comments that were similarly instrumental and only indirectly social, as they focussed on objects rather than the individuals that had posted them. This form of sociality resonates with Knorr-Cetina’s (1997) observations of increasingly distant and object-mediated forms of sociality in many Western societies. Similarly, Turkle (2011) argues that even though technologies appear to give us a sense of increasing connectedness, they have led to greater distance, allowing people to socialise in distant and risk-free ways. For example, Turkle (2011 p.xiv) describes how the randomly-paired conversations of Chatroulette take this to an extreme where “people are objectified and quickly discarded”. To some extent, instrumental interaction with DistractMe was similar, as the participants treated other users and their contributions as resources for quitting that were discarded when the nicotine withdrawal symptoms subsided. While this form of instrumental sociality was common, it was not the only form of sociality experienced in this study.

The participants experienced various forms of sociality by engaging with the social presences of others in the group. The simplest form was through being receptive to the presence of others. The trial of DistractMe showed that many participants browsed for social traces of other users and felt particularly connected with other users who posted their personal stories. The participants sometimes expressed empathy with what other users were experiencing, particularly with the difficulties of staying quit. Empathy is an important form of support in online communities (Preece 1999). Furthermore, being receptive represents an active form of engagement in an online community. This is often discussed as a form of legitimate peripheral participation (Lave and Wenger 1991), which can include engagement with the practices discussed in an online group, identification with its members, as well as a service as an audience for contributors (Antin and Cheshire 2010).

Another kind of sociality, described here as covert, occurred where participants experienced their actions as affecting others in the group but no representation of them was put forward to the group, genuine or symbolic. Like instrumental sociality, covert sociality was focused on objects and lacked personal information. However, covert sociality was indirectly social as it showed an explicit intent of influencing and possibly helping other users. This intent was visible through motivational messages that were posted to encourage quitting. In the case of participant 20, this also included the sharing of distractions to assist others, even though she had abandoned her own quit attempt and was smoking again. Covert sociality, as well as receptive and instrumental sociality, is a powerful form for ambivalent socialisers, because it allows participation in an online community without having to disclose personal information or expose any representation of presence. The absence of personal information about the behaviour change reduces potential peer pressure (Morris et al. 2010; Newman et al. 2011) and embarrassment due to failure to change (Purpura et al. 2011).

Despite the widespread reluctance of quitters to interact with each other online, some participants disclosed personal information via DistractMe. We described this as an act of expressive sociality. In their posts and comments, these users explicitly referenced themselves and their situation to the group. Self-disclosure can be significant in indicating a person’s commitment to the group (Ren et al. 2012). It helps to establish trust between members, foster relationships, and strengthen the group’s identity (Joinson and Paine 2007). Beyond that, self-disclosure has a variety of potential personal benefits, like increased mindfulness as well analytical distance from the challenges of behaviour change (Graham et al. 2009; Pennebaker and Seagal 1999). All of these personal and interpersonal benefits help to establish collaboration between group members, as illustrated through the reciprocal self-disclosure presented in the findings. The findings have also shown that self-disclosure was more common with tips than with distractions, because tips were typically textual and composed by the participants, whereas all distractions linked to existing online content.

Finally, we observed instances of what can be called mutual sociality, where participants explicitly referenced others, so explicitly acknowledging a shared interaction. Some participants wrote their posts to reach out to others and invited them to collaborate on the quit attempt. Furthermore, some participants responded through comments that offered explicit messages of support to specific other users. These interactions were rare, but they highlighted that even though the DistractMe app was designed to support interaction through and around objects, some users progressed to interactions between individuals in order to share social support. As in other online communities for health and behaviour change (Maloney-Krichmar and Preece 2005; Newman et al. 2011; Ploderer et al. 2013; Wang et al. 2012), the users exchanged social support through information, encouragement, validation and companionship. As with disclosure, the exchange of mutual support was more common through tips than distractions because they were composed by the participants and focussed on the issue of behaviour change.

6 Conclusions

This paper has presented the design and study of an object-centred smartphone app to encourage online interaction in the domain of behaviour change. This research has drawn inspiration from CSCW inquiries into the design and use of digital objects to shift the focus of support groups from the exchange of personal experiences with change to more incidental social interactions mediated by objects. The object-centred design approach offers significant possibilities to encourage interaction around behaviour change, because individuals are typically reluctant to engage with other users and seek support before they have changed successfully. The object-centred design presented in this paper addresses this issue by providing users with a non-social reason to enter a potentially social online space coupled with enhanced control over how much they want to engage with other users and what information they want to share about themselves.

An object-centred design approach may also be useful for domains outside the area of behaviour change where users may have the same contradictory desire yet reluctance to participate. For instance, research indicates that social media offer great opportunities for people who have lost their jobs, but individuals are often reluctant to use them in such a situation for fear of disclosing stress and anxiety (Burke and Kraut 2013). For these users, an online group designed around objects such as resources to obtain support and find a new job may offer a viable alternative.

Furthermore, the study presented here has provided a detailed account of the various forms of sociality afforded by an object-centred design. This account provides an important alternative to established models of online participation, such as the reader-to-leader framework (Preece and Shneiderman 2009). Rather than simply describing the objective extent of engagement within an online community, the dimension of sociality developed here (instrumental, receptive, covert, expressive and mutual) provides a more nuanced account of a user’s experience of the presence of the social group and his or her interactions with it. Given that many online communities are structured around objects, whether they be photos, videos, locations or similar, we hope that our account of object-centred sociality will aid other researchers in their analysis of online interaction.

Finally, managers of online communities may benefit from an object-centred approach in order to encourage interaction amongst its members. Rather than seeding online groups with questions or using competitions to encourage conversations between new members (Resnick et al. 2010), seeding of impersonal objects through community managers provides an opportunity to stimulate community experience and thereby attract users.

Hence, we hope that these examples and the work presented in this paper provide inspiration for other researchers and practitioners to search for suitable objects for their chosen domain, establish online groups around these objects, and exploit the opportunities offered by an object-centred approach to online interaction.